<b><i>Background:</i></b> Haemorrhagic and thrombotic complications are common in dialysis patients on warfarin; thus, accurate international normalized ratio (INR) monitoring is critical. For expediency and patient comfort, blood sampling from the haemodialysis access or circuit is commonly performed. Point-of-care (POC) INR machines allow both peripheral vein preservation and rapid results, yet are not validated in the haemodialysis population. <b><i>Methods:</i></b> A prospective cohort study in haemodialysis patients taking warfarin was undertaken. Three paired samples were drawn over a single session: peripheral blood INR, POC INR, and dialysis INR. Agreement using Bland-Altman analysis and correlation coefficients between the peripheral blood INR, haemodialysis INR, and POC INR were calculated. Inappropriate dosing decisions based on haemodialysis or POC INR were quantified. <b><i>Results:</i></b> Amongst 34 patients, agreement between the dialysis INR and peripheral blood INR was high, with the haemodialysis INR differing from the peripheral INR by <±0.2, 85.2% of the time. Correlation between the 2 methods was high (<i>r</i> = 0.914; <i>p</i> < 0.001). POC INR differed from peripheral INR values by <±0.2, 67.6% of the time, with less agreement at higher INR values. Dosing decisions were incongruent between the dialysis and peripheral INR in 6%, whilst the POC and peripheral INR disagreed in 26%. <b><i>Conclusions:</i></b> There was good agreement and correlation between the peripheral blood, haemodialysis access/circuit, and POC INR values. POC INR was less reliable at higher values, and dosing decisions differed from the peripheral INR in a quarter of cases.